add TLD validation support

lib/auto_linker/parser.ex

@@ -28,13 +28,15 @@ defmodule AutoLinker.Parser do
   # @invalid_url ~r/\.\.+/
   @invalid_url ~r/(\.\.+)|(^(\d+\.){1,2}\d+$)/
 
-  @match_url ~r{^[\w\.-]+(?:\.[\w\.-]+)+[\w\-\._~:/?#[\]@!\$&'\(\)\*\+,;=.]+$}
+  @match_url ~r{^[\w\.-]+(?<tld>\.[\w\.-]+)+[\w\-\._~:/?#[\]@!\$&'\(\)\*\+,;=.]+$}
-  @match_scheme ~r{^(?:http(s)?:\/\/)?[\w.-]+(?:\.[\w\.-]+)+[\w\-\._~:/?#[\]@!\$&'\(\)\*\+,;=.]+$}
+  @match_scheme ~r{^(?:http(s)?:\/\/)?[\w.-]+(?<tld>\.[\w\.-]+)+[\w\-\._~:/?#[\]@!\$&'\(\)\*\+,;=.]+$}
 
   @match_phone ~r"((?:x\d{2,7})|(?:(?:\+?1\s?(?:[.-]\s?)?)?(?:\(\s?(?:[2-9]1[02-9]|[2-9][02-8]1|[2-9][02-8][02-9])\s?\)|(?:[2-9]1[02-9]|[2-9][02-8]1|[2-9][02-8][02-9]))\s?(?:[.-]\s?)?)(?:[2-9]1[02-9]|[2-9][02-9]1|[2-9][02-9]{2})\s?(?:[.-]\s?)?(?:[0-9]{4}))"
 
   @default_opts ~w(url)a
 
+  @tlds "./priv/tlds.txt" |> File.read!() |> String.trim() |> String.split("\n")
+
   def parse(text, opts \\ %{})
   def parse(text, list) when is_list(list), do: parse(text, Enum.into(list, %{}))
 
@@ -183,7 +185,7 @@ defmodule AutoLinker.Parser do
     if Regex.match?(@invalid_url, buffer) do
       false
     else
-      Regex.match?(@match_scheme, buffer)
+      Regex.run(@match_scheme, buffer, capture: [:tld]) |> is_valid_tld?()
     end
   end
 
@@ -191,10 +193,13 @@ defmodule AutoLinker.Parser do
     if Regex.match?(@invalid_url, buffer) do
       false
     else
-      Regex.match?(@match_url, buffer)
+      Regex.run(@match_url, buffer, capture: [:tld]) |> is_valid_tld?()
     end
   end
 
+  def is_valid_tld?(["." <> tld]), do: tld in @tlds
+  def is_valid_tld?(_), do: false
+
   @doc false
   def match_phone(buffer) do
     case Regex.scan(@match_phone, buffer) do

test/auto_linker_test.exs

@@ -27,4 +27,58 @@ defmodule AutoLinkerTest do
              "<a href=\"#\" class=\"phone-number\" data-phone=\"8888888888\">888 888-8888</a>" <>
                "  <a href='a.com' class='auto-linker' target='_blank' rel='noopener noreferrer'>ab</a>"
   end
+
+  describe "TLDs" do
+    test "parse with scheme" do
+      text = "https://google.com"
+
+      expected =
+        "<a href='https://google.com' class='auto-linker' target='_blank' rel='noopener noreferrer'>google.com</a>"
+
+      assert AutoLinker.link(text, scheme: true) == expected
+    end
+
+    test "only existing TLDs with scheme" do
+      text = "this url https://google.foobar.blah11blah/ has invalid TLD"
+
+      expected = "this url https://google.foobar.blah11blah/ has invalid TLD"
+      assert AutoLinker.link(text, scheme: true) == expected
+
+      text = "this url https://google.foobar.com/ has valid TLD"
+
+      expected =
+        "this url <a href='https://google.foobar.com/' class='auto-linker' target='_blank' rel='noopener noreferrer'>google.foobar.com/</a> has valid TLD"
+
+      assert AutoLinker.link(text, scheme: true) == expected
+    end
+
+    test "only existing TLDs without scheme" do
+      text = "this url google.foobar.blah11blah/ has invalid TLD"
+      expected = "this url google.foobar.blah11blah/ has invalid TLD"
+      assert AutoLinker.link(text, scheme: false) == expected
+
+      text = "this url google.foobar.com/ has valid TLD"
+
+      expected =
+        "this url <a href='http://google.foobar.com/' class='auto-linker' target='_blank' rel='noopener noreferrer'>google.foobar.com/</a> has valid TLD"
+
+      assert AutoLinker.link(text, scheme: false) == expected
+    end
+
+    test "only existing TLDs with and without scheme" do
+      text = "this url http://google.foobar.com/ has valid TLD"
+
+      expected =
+        "this url <a href='http://google.foobar.com/' class='auto-linker' target='_blank' rel='noopener noreferrer'>google.foobar.com/</a> has valid TLD"
+
+      assert AutoLinker.link(text, scheme: true) == expected
+
+      text = "this url google.foobar.com/ has valid TLD"
+
+      expected =
+        "this url <a href='http://google.foobar.com/' class='auto-linker' target='_blank' rel='noopener noreferrer'>google.foobar.com/</a> has valid TLD"
+
+      assert AutoLinker.link(text, scheme: true) == expected
+    end
+  end
 end